Environment <ttul j'lannmx P Sofictv and Space WX, volume 16, pages 4N.1 MM

Spaces of surveillant simulation: new technologies, digital representations, and material geographies

Stephen Graham Centre for Urban Technology, Department of Town and Country Planning, University of Newcastle, Newcastle upon Tyne NK1 7RU, England; e-mail: s.d.n.graham«i nel.ac.uk Received 31 August 1996; in revised form 22 September 1997

Abstract. In this paper I analyse some emerging socioeconomic applications of information and communications technologies and explore how they support technological systems which increasingly blend surveillance with simulation. In the first part of the paper I explore the technological shifts supporting blended 'surveillant simulation' and review how the emerging links between surveillance, simulation, and material geographies have been addressed in recent debates on society, space, and cultural change. In the second part I go on to explore four examples in detail where widespread electronic surveillance systems are providing the captured data and images to produce electronic simulations of the 'real world' in near 'real time'—virtual banking, retailing and 'reality', crime control and electronic tagging, road transport telematics, and 'smart' utility systems. Attention is focused on how such simulations of the real world are then used to support new spatial practices based on the fme-grained allocation of goods and services, and intimate patterns of attempted social control, in real time, through the time-space fabric of material geographies. I conclude by analysing the implications of surveillant simulation for theories of technology, space, and place, for social polarisation in cities, and for considering opportunities for resisting the spatial practices of dominant organisations.

Introduction: surveillance, simulation, and the spaces of control Computerised surveillance and simulation have been subject to much recent debate within social theory and commentaries on society and space (for example see, Boycr, 1996; Drocge, 1997; Gregory, 1994; Lyon, 1994). In particular, the technological under­pinnings of the widening of computerised surveillance and simulation are now clearly established. Four can be identified. First, advanced telecommunications networks increasingly support the interlinkage of widening arrays of terminal equipment across geographical space into digital multimedia 'telematics' networks, able to deal with flows of digital data, sound, voice, and (increasingly) still and moving images (Lyon, 1994). Second, advances in computing technology mean that the powers of digital technological systems for processing, manipulating, transmitting, and storing data are increasing extremely rapidly This means that systems supporting new orders of magnitude of automated data capture, monitoring, and surveillance can be directly constructed to try and match vastly complex systems of social and economic behaviour across material spaces. Third, computers are, in turn, moving from being essentially 'data crunching' devices, to become sophisticated visualisation and simulation devices. New generations of geographical information systems (GISs), digital mapping, and virtual reality (VR) techniques are emerging where subjects are, to a greater or lesser extent, sensorily immersed in simulated environments through 'real-time' cybernetic monitoring and feedback (see Druckrey, 1994a). "The transition from solid models to digitally generated images has gone to completion in an astonishingly short time" (Stone, 1994, page 7). Fourth, the geographical foundations for the fme-grained moni­toring of the time - space dynamics of geographic spaces have been provided by rapid advances in georeferencing technologies such as satellite remote sensing, the global web of global positioning systems (GPS) satellites, and GISs (Pickles, 1995).

484 S Graham

Together, these broad technological trends mean that data surveillance techniques of unprecedented power and scope can now become spatially visualised and opera-tionalised through sophisticated GISs and, increasingly, VR and computer simulation technologies. Ultimately, technological enthusiasts and utopianists predict some simple collapse of the distinction between electronic surveillance and simulation as facsimile worlds become indistinguishable from 'real' ones. Here, fully immersive real-time virtual simulations are predicted that are so intimately connected to surveillance systems that they can be taken to be 'mirror worlds', 'software worlds in a box' (Gelerntner, 1991), 'intelligent environments', or 'virtual urban spaces' (Droege, 1997).

Gelerntner (1991, page 3), for example, predicted that software constructions, linked to a range of real-time surveillance inputs, would become such lifelike metaphors for the 'real' world that they were taken for "software models of some chunk of reality, some piece of the real world going on outside your window". In such mirror worlds, he writes, "oceans of information pour endlessly into the model (through a vast maze of software pipes and hoses); so much information that 'the model' can mimic the 'reality's' every move, moment-by-moment" (1991, page 3). Indeed, it is widely argued that, with current advances in GIS and VR systems, simulated facsimiles will become more and more like the real world. Jacobsen (1994, page 37), for example, believes that "the addition of virtual worlds to GIS will result in a hybrid technology, the living map, that enables users to naturally experience geospatial information and the world this information represents".

Unfortunately, however, approaches in social and cultural geography, and poststruc-turalist and political economic commentaries more generally, tend still to separate treatments of surveillance and simulation. In analysing the surveillance practices surrounding geodemographics, GIS, closed circuit television (CCTV), and other new information and communications technologies (ICTs) (Norris and Armstrong, 1997; Pickles, 1995), surveillance analyses have drawn on the work of Foucault (1977) on the disciplinary antecedents of modern societies (for example see, Philo, 1992; Squires, 1994) and on Jeremy Bentham's famous 18th century writings on his panopticon prison design (see Hannah, 1997). Analysts of simulation, meanwhile, have drawn on Baudrillard's (1983) notions of the postmodern shift towards a society dominated by technostrategic simulation (Der Derian, 1990), and by an increasingly pervasive hyperreality comprising ascending orders of simulacra. Recent simulation analyses have focused on the military 'cyberwar' simulators at the basis of the Gulf War, the growth of computer games and VR simulations, the operation of cybernetic consumption systems, and the 'them-ing'of commercial environments (see Baudrillard, 1983; Gottdeiner, 1997; Kellner, 1994; Soja, 1996). Other more technical debates, meanwhile, have explored the practical issues surrounding computerised simulations of everything from geographical systems (see Batty, 1996; 1997) and societal processes (Gilbert and Conte, 1994) to planetary landscapes, biological mechanisms, and the whole fabric of cosmological space - time (see Hall, 1993).

Some recent progress has been made in addressing the interlinkages between sur­veillance and simulation at the general level in recent poststructuralist and postmodern commentaries on society and space. Baudrillard himself was careful to note that the 'consumer simulacra' of geodemographic models were constructed through feedback created from the linkage of consumers into surveillance systems (Baudrillard, 1994; Sawchuck, 1994, page 109). Burrows (1997a; 1997b), following Davis (1990; 1992), has drawn on 'cyberpunk' readings of the city to build up theoretical lessons for analysing the links between cyberspace, simulation, and geographical change. Also, writers such as Boyer (1996), Davis (1992), Haraway (1991), Olalquiaga (1992), and Soja (1996) have started to explore the cybernetic relations linking surveillance, electronic representations, and

Spaces of surveillant simulation 485

materia! urban landscapes, where *cyborgian* feedback loops mean that the distinctions between surveillance and simulation become ever more blurred. Haraway asserts that the blurring of the body technology boundary within pervasive cybernetic systems leads to a "technological polls where maehinic-desires drive cybernetic systems by artificial instincts and recursive feedback loops" (Maraway, 1991, pages 149 - 181). To her, a cyborg world simultaneously generates grids of control over space- time and constructed simulations of embodied subjectivities.

Boyer, meanwhile, has argued that the undergridding of material cities with ICTs means that control now "acts like a sieve (a computer matrix) whose mesh transmutes from point to point, undulating and constantly at work. The code, not the norm, becomes the important device" (Boyer, 1996, page 18). Thus, the material geographies of cities and city systems becomes constituted through new nexus points between knowledge and power that transcend Foucault's architectonic elements to embrace combinations of "discourses and architectures, programs and mechanisms". These "seem to be dislocated from space, deeply hidden within the electronic matrices of a global computer network that connects all points in space and directs our lives from some ethereal 'other' location" (Boyer, 1996, page 163).

Davis's (1992, page 16) work on Los Angeles has helped to concretise such arguments. Me explicitly argues that Los Angeles's highly polarised social order and repressive sociospatial practices are now clearly founded on webs of high-tech surveillance, linked into both the urban simulations in 'cyberspace', and the extending material simulacra of theme-parked, consumption-driven, corporate landscapes (see Sorkin, 1992). Thus, Los Angeles:

"simulates or hallucinates itself in at least two senses. First, in the age of electronic culture and economy, the city redoubles itself through the complex architecture of its information and media networks... . Urban cyberspace—as the simulation of the city's information order—will be experienced as even more segregated, and devoid of true public space, than the traditional built city .... Second, social fantasy is increasingly embodied in simulacral landscapes—theme parks, 'historic districts' and malls—that arc partitioned off from the rest of the metropolis" (Davis, 1992, page 16). Further progress in exploring and theorising the interconnections between surveil­

lance and simulation has been made in the cultural studies field by Bogard in his recent book The Simulation of Surveillance (1996). Bogard's work is useful because it takes an holistic perspective of the complex interactions between surveillance and simulation which emerge with pervasive computerisation and the digitisation of grow­ing swathes of information. "It is simulation", writes Bogard (1996, page 9), "that is the key to explaining the direction that surveillance societies are taking today, a movement that is more about the perfection and totalization of existing surveillance technologies than some kind of radical break in their historical development".

Bogard suggests at the general level that contemporary Western societies are becoming "telematics societies", which are defined as "societies that aim to solve the problem of perceptual control at a distance through technologies for cutting the time transmission of information to zero" (1996, page 9). To Bogard, contemporary restruc­turing of advanced capitalist society is interwoven with the widespread application of military-standard computerised simulation technologies and command - control - com­munications webs. Echoing wider 'surveillant society' critiques (Davies, 1996), he argues that emerging webs of surveillance technologies tend to be woven closely within broader global-political-economic shifts which tend to operate to privilege large corporate interests over others; to minimise the barriers to corporate investment freedom and capital flow; to support the privatisation and/or liberalisation of public services,

486 S Graham

infrastructures, and institutions; to underpin the commodification of both the spaces of cities and the spaces of information; to accelerate the speed of flows and representa­tions; and to support the emergence of a global capitalist society, founded on dense webs of digitally interconnected control systems (Der Derian, 1990; Hewson, 1994; Schiller, 1996; 6 Tuathail and Luke, 1994).

Bogard argues that broad technological systems are increasingly being constructed which are both surveillant and simulating at the same time. Such trends now mean that GISs are increasingly being integrated with surveillance systems into a broad raft of other telematics technologies (home and transport telematics, VR, image databases, analogue and digital CCTV, DNA and biometric scanning, etc). Such integrated sur­veillance and simulation systems herald the emergence of 'hypercontrol' technologies designed to "push surveillance technologies to their absolute limit" and "to see every­thing capable of being seen, recording every fact capable of being recorded" (Bogard, 1996, page 4).

The central argument of The Simulation of Surveillance is that surveillance systems can now provide the data inputs necessary to develop electronic simulations of 'reality' to powerful organisations (the state and large firms). Such is the scale of the data that are continuously captured by these webs of surveillance, that the electronic signs and symbols that are continuously provided can be (reconstructed into real time, or near real-time, virtual simulations of reality which are then taken to be reality by dominant institutions. "The electronic signs/images of objects and events are taken for their 'real counterparts" (1996, page 3). The point "where the imaginary and real coincide" becomes so blurred that "the gap between virtual control and actual control disappears" (1996, page 9). Aggregated captured data then support computerised visualisation and simulation by allowing dynamic facsimiles of socioeconomic systems and material geographies to be dynamically constructed. Thus, as Yurick suggests, "whole nations, their economies, their peoples, their resources, their land, can be simulated and displayed as some electronic input/output device" (1985, page 3).(1)

Bogard's book can be accused of an overarching, essentialist, social determinism.(2)

It also fails to explore how practices of surveillant simulation become embodied in the production of grounded sociospatial relations. What is lacking in all of the above accounts, in fact, is the degree of close empirical detail necessary to unravel clearly the complex social practices and political economies through which surveillance and simulation become interlinked in the production of new material geographies. This is a problem because the growing nexus between systems of surveillance and those of simulation raises major questions about geographical change, social control, patterns of inclusion and exclusion, the development of visual culture, the construction of (and struggles over) subjectivities, and the spatial dynamics of the so-called 'information society'. How, for example, do extending and interconnecting ICT grids capture data that become aggregated into societal simulations? How are surveillance and simulation practices linked together within specific social contexts to feed back onto the produc­tion of material geographies? What social and spatial practices support the blending of (1) Many good illustrations of surveillant simulation come from the original source of most of the technological innovations involved—the military sphere (De Landa, 1991; Hillis, 1996). Here, as Robins (1996, page 55) puts it, "surveillance and simulation feed off each other. And surveillance and simulation technologies together feed into the control of a new generation of 'smart', vision-guided strike weapons". W The elaboration of sociotechnical 'hybrids' is 'messier', more dynamic, and more uncertain than the macrolevel political economic determinism inherent in Bogard's thesis. To adopt the terms of actor network theory (Bingham, 1996), efforts to 'enrol' surveillance and simulation systems to create social ordering, even by dominant institutions, are inevitably full of specificities, contingencies, subtleties, and what Pile and Thrift (1995, page 37) call "open-ended cosmologies".

Spaces of surveillant simulation 487

surveillance and simulation? What does the blending of surveillance with simulation mean for the ways we conceptualise space, place, and 'society-technology' relations? And what space exists for resisting the dominant surveillant simulation strategies of large corporate organisations, infrastructure providers, corrections agencies, and state power?

In the current paper I seek to begin exploring such questions in detail. I first examine four specific cases in which combined systems of surveillance and simulation are being woven into the fabric of material geographies and the geometries of social and spatial power relations: in retailing, banking, and home teleservices; in crime and social control; in road transportation, and in the utilities sector. Through these I aim to connect a broad perspective on the geographies of I d s with political-economic debates about surveil­lance, computerised simulation, and the socioeconomic restructuring of geographic space. I conclude by attempting to consider the implications of surveillance simulation for theories of space, place, and technology, for our understanding of dominant spatial practices and social exclusion, and for the possibilities of resistance.

Cybernetic consumption, surveillant simulation, and home teleservices Our first case centres on the emerging linkages, in the consumption sphere, between surveillance and simulation techniques in retailing and home teleservices (interactive cable television, phone, video on demand, etc). There are two levels to consider here: locational decisionmaking and investment targeting, and the use of new ICT-based consumption systems as automated survcillant-simulation networks. In locational deci­sionmaking and investment targeting, retailers and banks are increasingly integrating GISs and geodemographic targeting systems into store investment and disinvestment decisions (Goss, 1995). A growing industry exists supporting the sophisticated geo­demographic profiling of census tracts and households (Batey and Brown, 1995) with the now familiar litany of clusters—'affluent achievers', 'urban venturers', 'smokestack shift workers', 'grafiltied ghettos', 'bohemian melting pot', 'have-nots', etc. These arc precisely mapped onto urban spaces to aid locational decisionmaking, direct marketing, and customer targeting. Thus, the material geographies of cities and national spaces are, in effect, themselves translated into surveillant simulations, into fine-grained dynamic maps of consumption and spending potential, as the large geodemographic bureaus now attempt to capture more and more direct consumption information from store credit cards, credit bureaus, direct marketing campaigns, Internet responses, and the like and store this information in GIS-based 'data warehouses'. Goss (1995) outlines how such GIS-based systems become transformed from partial representations and simulations of reality to operate in effect as the trusted representation of reality, the basis for precise locational decisionmaking and profit-driven targeting within large retailers:

"the GDIS [Geodemographic Information System] is literally represented as a con­struction, a 'built environment' consistent, of course, with the architectonic meta­phors so pervasive in the discourse on information technology. This architectonic metaphor effectively gives substance to a language, reifying the binary code that represents information as an alternate world, literally a data 'structure' .... The abstract data structure is then anchored to a direct representation of reality, which leads to the conceit that the world of the GDIS is itself another reality .... Here is the perfect edifice for strategy, an ironic doubling of the interiority-exteriority relationship. A representation of the 'exteriority' of the world is interiorized on the computer. The world of the 'other' and its identity have been captured and contained on a spatial grid by the machine technology, where it can be system­atically observed and manipulated by the strategy and power on the other side of the screen" (Goss, 1995, pages 143-144).

488 S Graham

In the context of profit-driven restructuring strategies, the practices which result from the use of such surveillant simulations tend to involve the withdrawal of banks and retail outlets from poorer areas and the careful 'cherry picking' of the best locations from the sociospatial matrix of the city (Graham and Marvin, 1996). In the United Kingdom, for example, the main retail banks have used GIS techniques to select the least profitable branches as part of their withdrawal of 25% of all bank branches since 1979 (see Leyshon and Thrift, 1995).

But we also need to consider the possible broader role of surveillant-simulation systems in mediating access to increasingly cybernetic, tele-based consumption services, as technological trends seem likely to shift inexorably toward a consumption-driven 'information superhighway5 dominated by very large media and consumption conglom­erates (Mowshowvitz, 1996). As trends towards home-based consumption founded on telephone, the Internet, cable, and broadband home networks combine with the growing use of electronic cash (credit cards, smart cards, and 'cybercash' on the Internet), home-based shopping, banking, and consumption systems are emerging which precisely monitor, in real time, the consumption patterns of households. Rather than relying indirectly on aggregated or individual consumption data from the census and credit and information bureaus, as has been the practice in the postwar period (Pickles, 1995), these systems actually build up their own surveillant simulations of actual individual behaviour in (near) real time.

In the guise of supporting 'freedom of choice', individuals linked into such systems themselves engage in generating 'transactionally generated information' (TGI), so building up their own 'digital personas'—surveillant simulations for corporate use (Crawford, 1996). This raises questions about how self-generated surveillant simulations, built up covertly and geared to the needs of large corporations, are also involved in the construction and control of subjectivities and identities. Who, in other words, owns one's digital persona—the subject, the data bureau, or the transnational corporation? And what are the geographies surrounding the data flows, through which these surveil­lant simulations are continually constructed, updated, and refined? Stone writes that:

"out of the snail track of our passage through a world of myriad simultaneous opportunities for consumption, [monitoring corporations] build their own images of who we are, freed from the constraints of linearity of sense. Our doppelgangers are already free of the tyranny of localized subjectivity; they follow the geodesies of capital and of ideal citizenship. It's ourselves that haven't yet caught up" (1994, page 7). It is becoming clear that the practices surrounding TGI can have very real impacts

on the material geographies of opportunity, constraint, and restructuring. TGI is usually used for various forms of exogenous social control by credit bureaus and consumer service organisations undertaking restructuring based on so-called 'data warehousing'. TGI allows firms to track real-time consumption habits, preferences, and practices; to identify and blacklist high credit risk individuals, households, and areas; to target and deliver individually direct marketing campaigns; and to build up commodified informa­tion packages for reselling within the lucrative 'information marketplace' (Crawford, 1996; Mowshowvitz, 1996).

The apparently humble case of the supermarket customer loyalty card is currently a key route to personalised surveillance and cybernetic customer targeting in the UK and US food retailing industries. Whilst giving users access to discounts and free goods, such cards provide the technological infrastructure for mass, continuous surveillant simulation of customers by corporations, where firms had to rely on crude estimates previously. Each time a customer with a loyalty card buys goods, their card is 'swiped' through the 'electronic point of sale' (EPoS) terminal at the checkout. This allows an

Spaces of survcillani simulation 489

individual profile of consumption habits to be built up over time, which can then be aggregated to provide a real-time simulation of throughput through all stores. In turn, this can feed into ordering, logistics* storage, and supply-chain management. It also provides the raw material for 'mass customisation* and direct marketing, Masscy (1996, page 26) suggests that UK "retailers like Safeway and Teseo can now build detailed pictures of spending patterns based on data gleaned from loyalty card/swipes. Even­tually retailers will be able to target customers with offers specific to them—potentially setting special price details accessed by individuals using self-scanners'*. Such 'targeted incentive marketing' allows individualised marketing to support loyalty and profit margins. Thus, "if Mrs Smith buys a certain amount of groceries and the system knows from her purchases that she has two kids, the system should target her with child-related benefits. Those kinds of relationships are going to drive increasingly loyal, full-margin purchases" (Smith, 1990, pages 20-22, quoted in Sawchuck, 1994).

Practices of computer telephone integration (known as CTI) in customer telesales centres hint further at the complex geographies and subtle processes of inclusion and exclusion that surround on-line consumer systems. Such centres are now used by major retailers, banks, insurance companies, transport firms, airlines, and utilities. Telesale centres service regional, national, and even international markets from a single tech­nologically advanced node (through the use of free or local call phone tariffs linked into corporate telematics networks). By automatically surveilling the source of incoming tele­phone calls, through a system known as call line identification (CLI), and linking this number into customer databases, such systems now allow callers to be sifted according to how 'good' a customer they arc. In effect, the surveillance of the caller is automat­ically linked to a simulation of all known customers, to allow customers to be treated differentially. Thus, UK utilities are already able to answer the calls of 'good' customers (those that have paid their bills promptly) before 'bad'customers (those with a history of default who are queued), without either being aware that their prompt or slow service is, directly shaped by automated surveillance systems linked to computerised databases^

My final example, that of home-based on-line consumption services, shows how automated cybernetic consumption systems can automatically generate and revise simulations of consumer potential, generating cybernetic feedback loops of targeted consumption, advertising, and media flow. For example, the much-vaunted experi­ments in interactive broadband home telematics, such as the Time Warner interactive television system at Maitland, Florida, are experimental precursors to the much wider rollout of highly capable home media and consumption systems, which are intrinsically based on building up surveillant simulations of consumers' behaviour (Burstein and Kline, 1995). The telecommunications company, Bell Atlantic, is developing a computer system linked to 'video on demand' (VOD) which will monitor the movies watched in each household, target advertising and product offers to households that are most related to the type of movies viewed, and aggregate all these data into a cybernetically improved simulation of the whole network population. Alliances of banks, satellite firms, retailers, distributors, and content providers are now designing 'personalised multimedia architectures' so that targeted advertising, consumption with what is called 'Tine granularity of control", can be constructed at an individual level across advanced industries markets. Dedrick (1995, page 43) gives an example of how this might work within a single family:

"For example, Dad, a male, age 40 to 50, earning $70,000 plus annually, might see part of a Mt. FunSki ski resort ad about booking a reservation, along with a list of fun things to do. However, the consumer's son, male, age 12 to 17, interested in girls, moguls, and hot tubs, might see a presentation based on [a different story of] the "fun things" that Mt. FunSki has to offer."

490 S Graham

In the longer term, it seems likely that home consumption systems will develop VR applications, accessed via interactive broadband home networks. VR, with its simulated 'environments' constructed through software and its constructed human actors strapped into gloves, masks, and exoskeletal devices, is perhaps the purest example of surveillant simulation. The total psychological immersion of VR means that its every aspect derives its functionality from visual simulation, integrated closely with real-time sur­veillance and cybernetic feedback loops, which link movement and thought to changes in the software simulation (Druckrey, 1994a). But, in reality, given the huge corporate resources now being devoted to them, such virtual environments seem likely to be defined largely for commercial ends. "As the system of technology expands to dominate the regulation of the external world", suggests Druckrey (1994a, page 9), "it also con­tracts and increasingly penetrates the internal world. The body is unquestionably the next frontier—the body, and then cognition".

The construction of VR-based consumption spaces will intensify trends towards real-time surveillant simulation and network arrangements based on ICT networks, which are tailored towards the needs of more affluent consumers. The majority of such spaces seem likely to be based on simulations which group together immersive con­sumer telematics services based on analogies of idealised cities or malls. They are likely to develop as three-dimensional versions of the simulated 'web cities' now common on the World Wide Web. Sheth and Sisodia (1993), for example, predict the development of 'information malls'—whole 'virtual' service spaces accessed from the home offering personalised retailing, teleworking, medical, education, financial, travel, messaging, video conferencing, and real-time news and information services (Sheth and Sisodia, 1993, page 382).

Surveillant simulation as social control: CCTV, tagging, tracking, and biometric scanning Our second example centres on the emerging links between surveillant-simulation technologies and crime and social control initiatives, particularly in cities. In these surveillant simulations, human subjects can effectively be reduced to their time-space electronic trails or representations, as their movements and behaviour are logged, tracked, and, increasingly, mapped via systems linking CCTV, computerised tracking systems, GISs, and mobile and fixed phone networks. The rapid extension of such technologies across geographic space means that "a person going about his or her daily routine may be under watch for virtually the entire time spent outside the house" (Squires, 1994, page 396). Druckrey (1994b, page 15) notes "the increasingly invisible dispersal of electronic tracking" technologies. Through wide-area systems covering whole cities, regions, nations, and international transport networks, the behaviour of human subjects may increasingly become aggregated into detailed time - space surveil­lant simulations, which offer radically new possibilities for tracking, social control, and targeted marketing.

Public wide-area CCTV camera systems in the United Kingdom are a good example of the emergence of surveillant simulation as (attempted) social control. More than 200 CCTV schemes in public places now exist in the United Kingdom, most of which use analogue video technology backed up by radio, telephone, and photographs of target subjects (Graham et al, 1996). Virtually every sizable urban settlement in Britain now has public CCTV; increasingly, systems are also spreading to cover residential areas.

Already, evidence is building up that; through CCTV, people and behaviours seen not to 'belong' in the increasingly commercialised and privately managed consumption spaces of British cities tend to experience especially close scrutiny. Norris and Armstrong (1997, page 4) have shown that operators of current, nondigital CCTV

Spaces of survcillnnt .simulation 491

"selectively target those social groups they believe most likely to be deviant. This leads to an over-representation of men, particularly if they are young or black1'. They show that CCTV control rooms are ridden with racism and sexism, and certain types of young people are targeted with socially constructed suspicion and are labelled as 'tocrags', 'scumbags', 'yobs', 'scraphcads', 4Big Issue scum*, 'homeless low-life', or 'drug-dealing scrolcs', and are scrutinised, followed, and harassed. Malign intent is directly equated with appearance, youth, clothes, and posture. Thus, operators are already imposing a 'normative space-time ecology' on the city by stipulating who 'belongs' where and when, and treating everything else as a suspicious 'other' to be disciplined, scrutinised, controlled. "For operators the normal ecology of an area is also a 'norma­tive ecology' and thus people who don't belong are treated as 'other' and subject to treatment as such" (Norris and Armstrong, 1997, page 43).

Currently, then, the surveillance within CCTV is not linked to electronic simulation; rather, the human eye and brain of the operator (with all its subjectivity and discretion), linked to police records and photographs, becomes the route through which CCTV imaging is translated into disciplinary social action and attempted control. But tech­nological developments towards the digitalisation of CCTV, and its automatic linkage with databases, seem likely to lead to much higher degrees of automation and a much greater reliance of linked survcillant-simulation techniques. Emerging, digital CCTV systems arc algorithmically programmed to scan for certain 'unusual' events or targeted individuals or vehicles, so withdrawing opportunities for human discretion in tracking and monitoring individuals. Digital CCTV will allow real-time time-space searching for specific events to occur as well as retrospective digital searching aimed at correlating behaviour patterns with patterns of crime (Norris et al, 1996).

Early examples of digital algorithmic CCTV applications arc already emerging. Certain UK rail stations now have 'smart' CCTV which automatically warns when specific crowd densities are met on platforms. On the Ml motorway in Northampton­shire an automated speed violation detection camera attempts to deter drivers from speeding. When a speeding car is detected its number plate is automatically scanned and the offence is put up on a large neon sign, Digital records are sent electronically to the local police headquarters of the car number plate and vehicle identification, who then communicate with the Driving Licensing Authority and policy computer opera­tors, who, in turn, automatically send a fine to the owner's address in the mail. The City of London now has an 'intelligent screen monitoring' algorithmic system for automated surveillance of its 'ring of steel5 antiterrorist cordon. Here, a stationary vehicle triggers an alarm in the control room as does a car heading down the street in the 'wrong' direction (Norris et al, 1996).

With respect to face recognition, a computerised database of 6000 football 'hooli­gans' has already been constructed to be sent on-line to police monitoring stadia CCTV systems. Antinuclear, antidefence, and green protesters, in addition to 'travellers', have also been scrutinised in this way. Sydney airport will soon introduce a system which scans automatically and covertly for known illegal immigrants entering immigration (Norris et al, 1996). In a new experimental project, British Telecom (BT) is working with the Massachusetts Institute of Technology (MIT) and the major British retailer Marks and Spencer on a digital image and television-based computer system known as 'Photobook' to be installed in its stores (McKie, 1994). Real-time cameras, linked to image databases of the faces of convicted shoplifters, will alert security staff of the presence of these shoplifters in their stores through advanced face recognition soft­ware. Accuracy is said to be "greater than 90%" (McKie, 1994). As Norris et al put it, such automatic surveillance and disciplining systems mean that "the complex moral

492 S Graham

calculus of policing is reduced to a mathematical formula. There is no discussion, no negotiation, no compassion and no empathy" (1996, page 16).

In the long run, BT anticipate major new telecommunications markets through operating such systems. For example, "all commercial outlets in a town could be linked and an alarm be set off the moment a person who has been seen shoplifting in one store enters another" (McKie, 1994). When backed by national digitised face prints of the type now being developed by the UK driver and vehicle licensing agency and passport authorities, the potential for national face-recognition and monitoring systems in the United Kingdom, operating through extending grids of CCTV, seems a lot more than some paranoid dystopia (Davies, 1996). Norris and Armstrong (1997, page 16) are very serious when they argue that the growing linkage of databases with widening CCTV grids means that "in approximately 20 years time, it will be possible for a national database to track the movements of our 'digital selves' around the country. Such a database could come from the new drivers license with a digital face record, or an identity card".

In the United States, the control and surveillance capabilities of ICTs are already being widely explored as tools for new methods of time - space tracking in cities. Even by 1991, over 4.3 million Americans had been under 'correctional supervision' at home (Gowdy, 1994). Anklet transponders, linked to telephone modems, provide continuous monitoring of the location of offenders. Newer 'smart' systems promise a much more fine-grained and tailored control over the behaviour of offenders. For example, in a store, the "arrival of an ankleted shoplifter would set off a silent alarm, and the system would identify the offender to the store management" (Winckler, 1991, page 35). When linked to wider urban surveillance systems through city wide radio networks—which will be available by 2000—the movements of all ankleted offenders could be correlated with the incidence of crime, in time and space, to help in conviction. "Every place the offender went—and the time he or she was there—would be recorded and compiled and could then be cross-indexed against known crime scenes and times" (Winckler, 1991, page 35). Thus, within this GIS-based surveillance simulation system, the 24-hour electronic tracks of individuals could be correlated with time - space patterns of crime incidence to underpin unprecedentedly fine-grained mechanisms for attempted social control.

Such initiatives are part of a much broader trend towards integrated real-time spatial simulations in policing. In Central Scotland, for example, the MEMEX system has been adopted by the regional police force to integrate all incoming information coming into the police into a seamless computerised simulation of the time - space situation of their area. Phone conversations, reports, tip-offs, hunches, consumer and social security databases, crime data, phone bugging, audio, video and pictures, and data communica­tions are inputted into a seamless GIS, allowing a relational simulation of the time -space choreography of the area to be used in investigation and monitoring by the whole force. The Chief Constable states: "what do we class as intelligence in my new system in the force? Everything! The whole vast range of information that comes into the possession of a police force during a twenty four hour period will go on to my corporate database. Everything that every person and vehicle is associated with" (quoted in Norris and Armstrong, 1997, page 91).

With electronic tagging already being applied to pets, babies, and cars, and with a range of biometric scanning devices emerging (smart cards, fingerprints, iris scans, hand geometry scans, thumb scans, voice recognition, DNA testing, and digitised facial recognition), the potential for citywide, national, even international systems of intimate monitoring and control, classifying individuals into the spaces and times where they 'belong', is clearly of great concern (Davies, 1996, page 62). Biometric signatures, linked

Spaces of surveillant simulation 493

to database simulation, are already widely embedded in algorithmic systems which automatically control physical access and movement. Davis (1997) reports that iris-scanning automatic teller machines (ATMs) were operational in Japan from 1997; inmate retina scanning is in operation in Cook County, Illinois to control prisoner movements; Connecticut and Pennsylvania are practising digital finger scanning to reduce welfare fraud; frequent travellers between Canada and Montana now use automated voice recognition for speedier throughput; and hand geometry scans are now made of immigrants in San Francisco to check for illegal immigrants. Moreover, a national DNA database now operates in the United Kingdom for tracking criminals and in 1994 the Labour Party even considered the introduction of a national DNA database of all citizens for such purposes (Davies, 1996). Commercial DNA databases are raising the spectre of gene tests to underpin surveillance systems in the health insurance industry, excluding people with genetic 'abnormalities' and high disease risk (Weiss, 1996). "The prospects of interoperable, even networked database raises a fright­ening spectre", writes Davis (1997, page 174). "Our body parts and prints could soon be bought and sold like Social security numbers by direct marketers, government clerks, or medical providers. One 'harmless' little retina scan, some ophthalmologists warn, could indicate that a person has AIDS or abuses drugs".

Such examples hint at the kinds of technological systems that may emerge with the blending of biotechnology, genetic scanning, biomctric surveillance, and ICTs. There? seems to be little doubt that the ongoing digitisation of the human genome will bring-new potentials to biomctric surveillance (Wilkie, 1996). Such trends need to be viewed as part of the shift to what Haraway calls a cyborg (cybernetic organism) world, which is ultimately about the "final imposition of a grid of control on the planet" (Haraway, 1991, page 153; quoted in Gregory, 1994, page 162). BT have even speculated that computer chips would be so powerful by 2025, that they could be biologically inte­grated into people's minds, to record everything they saw and everything they thought, throughout their entire lives—an alleged boost to personal communication. Although this might sound like little more than science fiction fantasy, it seems likely that the collapse of the boundaries between information technologies and biotechnologies will mean that techniques of surveillant simulation will diffuse into the human body and mind itself, creating integrated "information organisms" (Davies, 1996, page 143). "Bodies that can be surveilled can be informated", writes Bogard (1996, page 63), "and bodies that can be informated can be reduced to codes". Druckrey (1994b, page 5) suggests that such surveillant simulation will mean that "encoding rather than identity could become the signifier of the self, an informatics of domination made possible only by the computer".

Surveillant simulation and differential power over space: road transport informatics My third case is that of road transport informatics (RTI). The control capabilities that new surveillant-simulation technologies bring are of central importance here, in sup­porting a shift from 'dead' public electromechanical highways to 'smart', digitally controlled, and, increasingly, privatised highways (Graham and Marvin, 1996). In the face of urban transport crises and the political and financial constraints on new urban road building, virtual electronic networks of automated sensors, CCTV, tracking and charging devices, computers, and GISs, are being laid over established road transport networks, helping to undermine their 'natural monopoly' characteristics, so allowing private firms to operate them profitably (Robins and Hepworth, 1988). Road networks, with all their complexity of flows and pattern, increasingly become surveillant simula­tions supporting new practices of commodification, control, and exclusion that provide the basis for strategies which differentiate groups according to the power over space they are seen to warrant within the new urban political economy.

494 S Graham

Within RTI systems, people and their vehicles are reduced to their moving image and signature within surveillant-simulation systems. Through the use of such technol­ogies to turn public road space into a private commodity, people become disaggregated according to how much power over space and mobility they warrant by the use of the price mechanism. This allows a dualisation process to occur between the poor and marginalised and the mobile, corporate elite groups who can access the commodified 'premium' road time - space and who are saturated with travel opportunities and access to mobile and fixed telecommunications networks (Graham and Marvin, 1996, chapter 6).

An excellent example of how surveillant simulation becomes implicated in the con­struction of new, dualised urban highway networks, and systems of differential power over space can be found in the construction of a new, private, commodified highway network (number 407) around Toronto. Built to ease congestion on the world's busiest highway, to which it runs parallel, in-car transponders will automatically charge all users of the highway around $1 per 11 km trip, without requiring them to stop. Tariffs will vary automatically, to peak around the busiest periods, to ensure that use of the highway never exceeds predefined limits, so allowing moving traffic to be guaranteed. Cars without transponders will be automatically photographed and their owners tracked. By the year 2000 over $100 million per year is expected in tolls, and speed limits may even be higher on the private highway than for other state highways. The consortium which built the road is now selling off all the key development sites along it to the highest bidder, for malls, affluent neighbourhoods, business parks, and logistics, creating, in effect, a second-tier land-use - transportation system for the elite interests in Toronto.

Electronic road systems directly supply the data that can be processed into real-life simulations of the entire road traffic time-space movement patterns of cities. Toll systems, which directly track known individuals to be billed later, now operate in Boston, New York, Florida, and California. Such trends may prefigure truly remark­able surveillant simulations, particularly of cities and high-traffic urban corridors. With the likely extension of toll systems over whole urban highway networks, and the rapid improvements in data capture, Garfinkel argues that it is not fanciful to imagine that, within a matter of a few years, a single digital tape could "store the record of an entire day's commute in a major metropolitan area, including 10 million hour-long trips, recoding every car's position every second (a total of 576 million bytes) and a photo­graph of every car every five minutes (a total of 3.6 Gigabytes)" (Garfinkel, 1996).

Using such tapes to build up real-time surveillant simulations of city transport patterns might then, in turn, be used to support precise, profit-driven planning by infrastructure corporations; the detailed analysis of people's driving habits; the track­ing of criminals or individuals; the enforcing of payment; covert police surveillance; and the reselling of the valuable information as a marketing tool to a whole range of information bureaus and direct marketing firms. Already, many cash-strapped munici­palities are considering how to capture such data for resale in the lucrative information marketplace. With single tapes allowing, in effect, total knowledge of the time - space movements of entire Metropolitan regions or urban corridors, in addition to the TGI data trails discussed above, marketing bureaux would be able to build up complete digital simulations of lifestyles with total precision, allowing ever more segmented and cybernetic direct marketing practices to be built up.

'Business process reengineering' and surveillant simulation in essential services: the case of the UK utilities Our final case centres on the UK utility industry, where GIS is rapidly being applied to allow utility corporations to take advantage of the newly liberalised regulatory situa­tion, to make utility planning and customer targeting closely resemble that in retailing

Spaces of .survcillant simulation 495

and financial services noted above (Guy et al, 1997). In the UK utility sector, national and regional public monopolies have recently been transformed into profit-hungry shareholder-driven enterprises. Utility firms are ^engineering* their structures and engaging in ever more global mergers, acquisitions, and strategic alliances within the internationalising energy, telecommunication, and water industries. Increasingly, they are engaging in survcillant simulation to support the competitive *chcrry picking' of lucrative market segments within cities whilst withdrawing from unprofitable social and spatial commitments (Graham and Marvin, 1994; 1996).

Three trends are noticeable here. First, GIS and computer-aided design (CAD) is providing the basis for the sophisticated mapping and planning of hidden utility infrastructures onto simulation systems. As in the retailing sector discussed above, geodemographic profiling is being used to support improved infrastructure planning and direct marketing, where, traditionally, utilities had very broad-brush, and generally relatively homogenous, approaches to rolling out and selling their services (see, for example, Saced, 1993). Mercury telecommunications, for example, have installed a GIS that allows customer locations to be ovcrlayccl with network maps so that it can "calculate whether the revenues it gets from its infrastructures justifies the costs, allowing it to question all the relationships it has with customers" (Bray, 1995, page 6), in order to maximise profitability and minimise loss-making investments. Corporate ICT systems are linking planning, operations, billing, and customer targeting systems together to allow for careful targeting of profitable areas and users whilst companies withdraw or ease out of costly social commitments (Graham and Marvin, 1994). Customer loyalty schemes are being developed to foster intimate knowledge over lucrative consumers.

Second, 'dead* utility meters are increasingly being transformed by ICTs into inter­active 'smart' systems for surveying and simulating the real-time household behaviour of lucrative market segments (Guy and Marvin, 1995). In the utilities sector, smart metering technologies, which link computerised meters via telecommunications to central control computers, now allow affluent consumers to choose their electricity, gas, or even water supplier by programming their own meter. Two-way communication then alerts their chosen utility that they now have a new customer, whose resource con­sumption can be remotely monitored and debited—avoiding the huge transaction costs associated with manual meter reading. In processes very similar to the commodifica-tion of road transport, telematics can therefore help to open up 'monopoly' networks to competitive supply, even in the domestic sector, without the need for replicating new energy systems. Such smart metering technologies then enable utilities to build up detailed information on household consumption patterns that goes beyond total use of services. Smart electricity meters can identify what electrical appliances a household owns, how often they are used and for how long. It is not even clear who 'owns' this type of information, but it opens up the potential for new forms of surveillance in the home. Here, households are effectively reduced to their telemetry signal representations. Real time residential powerline surveillance (RRPLS) can be used to provide an intimate profile of households' appliance usage (for example, it can track houses, where high-powered lights are being used to grow marijuana), which allows insurance companies to calculate the risks associated with different appliance users, utility firms to select carefully target markets, and large profits to be made in the information resale market (Crawford, 1996).

Third, at the other end of the utility market, households in marginalised communi­ties will, increasingly be rendered: invisible to the utility, because they are classified as high-cost or low-income customers. Such customers either fall off the networks because of the increased costs they face (through the rebalancing of tariffs), or are forced to use digitally encoded prepayment meters, based on smart cards which are 'topped up' at

496 S Graham

charging points, for accessing even basic utility services. In the United Kingdom, several million households currently use such meters for electricity consumption, and water prepayment is on the way.

Concluding discussion In this paper I have attempted to trace empirically some of the complex emerging intersections between new technologies, practices of surveillance and simulation, and the production and reconfiguration of material geographies. Its wide breadth of coverage has allowed broad parallel trends to be identified across many 'sectors', from retailing, banking, and home teleservices, through crime control, tagging, and tracking, to the (re)commodification of road transport and utility infrastructures.

It is clear that the importance of trends towards the widespread application of surveillant-simulation techniques is that they support increasingly coordinated, exten­sive, and comprehensive systems of surveillance, sociospatial targeting, and (attempted) social control. Technological developments linking surveillance with societal simulation, and the increasing horizontal coordination between dossiers and sites of surveillance (credit bureaux, banks, retailers, utilities, state and correctional agencies), seems likely to prefigure a rapid intensification of coordinated, comprehensive surveillance. The broad result of these shifts is the development of sociotechnical 'systems' which attempt to monitor, control, and guide social processes with unprecedented precision and power, and which are virtually invisible and formally unregulated.

Above all, it is becoming more and more difficult to escape, to lift a phrase from Latour (1993, page 121), from the "skein" of sociotechnological networks that undergrid the apparatus of surveillant-simulation. With their widening horizons of automated data capture, their continuous feeding of societal simulations, and their (near) instanta­neous geographical reach, it would seem that "we are in a generalized crisis in relation to all environments of enclosure .... Societies of control are in the process of replacing disciplinary societies" (Deleuze, 1988, page 4). As Lilly puts it, "the new surveillance transcends distance, darkness, and physical barriers. It can look through doors, suitcases and far into the inner intellectual, emotional, and physical regions of the individual" (1990, page 500).

Clearly the breadth of this paper raises many key questions for theories of socio­spatial change; for relations between the 'global' and 'local'; for the links between the contingent subjectivities of technological embodiment and broader political economic processes; and for the links between 'virtual' and 'real' worlds. But, within the space constraints here, I would like just to address four key questions.

First, there is the question of what the trends outlined might amount to, at the macrolevel and in the medium term. Do trends towards surveillant simulation neces­sarily prefigure some wholesale shift towards societies of dystopian social control and segmentation [as is so often implied in cyberpunk science fiction and critical social theory (see Burrows, 1997a; 1997b)]? Certainly, there is little doubt that the techniques of surveillant-simulation mean that "previously unconnected surveillance threads are now woven into gigantic tapestries of information" (Lilly, 1990, page 491). Within and across geographic space, electronic surveillant simulations are increasingly being constructed in the corporate and state spheres to support decisionmaking, business restructuring deci­sions, and the development of further iterations of surveillance by service organisations. Although relatively crude at present, rapid advances in processing power and computer vision technologies seem likely to enhance dramatically their capabilities within a short time-period. Within the context of political economies apparently dominated by an increasingly profit-driven, liberalised/privatised and internationalising corporate envi­ronment, surveillant-simulation systems are emerging as crucial techniques for bolstering

Spaces of surveillitnt simulation 497

profitability, flexibility* and responsiveness amongst dominant corporate, correctional, and state organisations.

For retailers, banks, and utilities, for example, survcillant-simulation systems are increasingly being woven into processes of 'business process rccnginecring*, service restructuring, and social targeting. This makes it possible to drive service plans and the rollout of investment across cities according to tight geodemographic targeting criteria. The time-space movements of subjects, their habits, desires, and preferences, become parodied within what Lyon calls

"'complementary selves' for those subjects; the sum, as it were, of their transactions. New individuals are created who bear the same names but who are digitally shorn of their human ambiguities and whose personalities are built artificially from matched data. Artificial they may be, but these computer 'selves' have a part to play in determining the life-chances of their human namesakes" (Lyon, 1994, page 71). Indeed, as cybernetic loops monitoring citizen behaviour become more sophisti­

cated (through retailers' customer information collection, mail order, consumer credit, profiling agencies, home ICT systems, road transport informatics, wide-area CCTV, etc), it is increasingly becoming possible to replace aggregate geodemographic spatial data sets (say, at postcode or census tract level) with individualised sets based on actual subject behaviour or consumption trails. Thus, at first sight, simulations of geographic space, which wc can consider to be examples of Lefcbvre's (1984) 'representations of space*—of cities, city systems, regions, nations, and international spaces—threaten to emerge which ever more closely resemble panoptic real-time simulations (the best example here being CCTV). Such panoptic and cybernetic networks increasingly start to resemble ' the command-control and communications webs already developed in the military. In the consumption field, the process of targeting reaches its limit as service enter­prises attempt to compete for market share within increasingly liberalised markets (whilst, of course, gradually easing out of less profitable commitments or obligations covering poorer groups and areas).

It is also important to stress that surveillant-simulation technologies are being developed and applied, within the context of a strong supply push, from an increasingly internationalised complex of media, ICT, and 'correctional' industries. What Lilley and Knapper (1993) call the "corrections commercial" complex—the fast-growing complex of security, military, and prison corporations who are, after the Cold War, attempting to colonise civil markets—are also key players in this supply-side push. They are being further supported by the broader debates about the supposedly world-improving momentum of the "information superhighway", the imperative to apply telematics uncritically to every aspect of civil life, and the pervasive crisis of public confidence in home, street, and transport security.

Increasingly, then, disciplinary control across material geographies comes to rely not just on the Foucauldian array of physical structures, disciplinary controls, and urban planning practices (see Driver, 1984), but on pervasive webs of electronic systems which assert disciplinary control by "distributing bodies/uses in space, allocating each individual/function to a cellular partition, creating an efficient machine out of its analytical spatial arrangement" (Boyer, 1996, page 17). As Virilio (1987, page 16) suggests, cities are shifting from a state where physical barriers and walls controlled access and 'belonging' to a state where "the rites of passage are no longer intermittent—they have become immanent" and are woven as automatic cybernetic systems into the urban fabric. New intensified technical systems linking knowledge, oppression, and the built environments of cities become constructed (see Laws, 1994). "We are obliged", argues Boyer, "to pay close attention to the links and nodes that interlace reality and appear­ances, illusions and symptoms, images and models" (1996, page 50).

498 S Graham

Thus, webs of simulated surveillance systems become woven into supporting and constructing the fabric of 'real' material geographies, just as the 'real' landscapes of cities and regions are themselves being peppered with 'theme park' simulations—the themed, hyperreal, commodified enclosures built by corporate leisure and media capital (Sorkin, 1992). Burrows (1997b, page 242) notes that "the digitised hyper-reality (of cyberspace) connects in various ways with the technological reality of the street, not least in the way in which the sociogeography of the digitised city mirrors that of the built city". Bleecker (1994) has even argued that changes of 'real' material geographies, particularly cities, seem to echo, and coevolve with, the engineering of virtual reality simulation games such as the screen-based SimCity 2000™ and its psychologically immersive descendants. Both in 'real' cities and in urban simulation games, Bleecker suggests, god-like operators can access even more comprehensive information sets about the processes of urban life, allowing them to create increasingly realistic simula­tions of the real-time dynamics of urban areas. The development of cities is then 'programmed' through computerised models to see the results visualise before the eyes through direct cybernetic feedback—the 'city as laboratory'. In both virtual reality and urban reality, Bleecker has argued that the result seems to be the emergence of a increasingly dystopian urban reality, approximating more and more closely the many postmodern, Blade-Runner-style scenarios of total corporate power, absolute surveil­lance, and urban decay.

Of course, the social world, with all its messy confusion of social practices, cannot be simply 'programmed' by computer. But when such techniques back up dominant spatial practices, as Davis (1992) has suggested in Los Angeles, we might see the emergence of urban landscapes made up of many superimposed layers of surveillant simulation and their associated practices. Each layer might have its own finer and finer mosaic of sociospatial grids; its own embedded assumptions and criteria for allocating and withdrawing services or access; its own systems for specifying and normalising boundary enforcement, through electronically defining the 'acceptable' presence of individuals in different urban 'cellular' time-space; and its own cybernetic loops of system feedback, within which systems of surveillance become ever more integrated into systems of simulation. As people leave a stream of digital tracks through their daily lives, their electronic personas become embedded into a web of surveillant-simulation systems—"each of us will become increasingly isolated in our own separate technologi­cal enclosure or cell" (Crawford, 1996, page 73).

Our second question concerns the intersection between surveillant simulation and social polarisation. In the current context, page 73, technological webs of surveillant simulation seem very likely to support processes of sociospatial polarisation, partic­ularly within and between cities. There are signs that surveillant-simulation techniques are being configured towards the protection, fortressing, and targeting of affluent consumer neighbourhoods and corporate districts (which are intensely linked into high-level transport and telecommunications infrastructures and become the focus for competitive cherry picking and retail and service locations), and to the exclusion, enforcement, and control of the groups and areas that are marginalised by labour-market and welfare restructuring (Davis, 1992; Graham and Marvin, 1996).

Thus, in US cities at least, surveillant-simulation techniques seem to be an essential prop to the separation of what Boyer (1996, page 25) calls the "figured city"—the well-designed, self-enclosed, protected, and closely-integrated nodes for work, consumption, affluent housing, leisure, and movement—and the "disfigured city"—the abandoned, neglected, interstitial spaces for the homeless and poor which "remains unimageable and forgotten and therefore invisible and excluded". In the context of the spatial hyper-fragmentation of urban politics—as epitomised in the USA by business improvement

Spaces of survetllant simulation 499

districts, drug enforcement zones, and homelessness management areas of New York and Los Angeles- Buyer points out that "the problems of drugs, hopelessness, or security are wrongly treated as boundary maintenance issues, where in fact they permeate the entire city'* (Dover, 1996, page 25; see also Flusty, 1994; Sojiu 1996).

Surveillant-simulation techniques add greatly to the power and precision of such boundary maintenance, especially when backed up by consumerist geociemographic targeting,'zero tolerance* strategies of policing, and welfare restructuring. Although we cannot assume the perfection or omnipotence of surveillance, we can suggest that surveillant-simulation practices are very likely to be operated "to ensure that in pro­tected zones defensive actions might be taken in response to invasion" (Dunn, 1993, page 186). Flusty, in his pioneering topologies of fortressed space in LA, writes that "under the auspices of municipal authorities, Los Angeles as a whole is well on the way to becoming jittery space", which he defines as "space which cannot be utilized unobserved" through combinations of CCTV, urban design and planning practices, helicopter surveillance, corporate databases, and the high-tech communications webs of correctional institutions (1994, page 37). To Flusty:

"Los Angeles is undergoing the invention and installation, component by component, of a physical infrastructure engendering electronically linked islands of privilege embedded in a police state matrix. If left unchecked, this trend may be linearly extra­polated into a worst case composite of hard boundaries, checkpoints and omnipresent surveillance. Los Angeles will become a city consisting of numerous fortified cores of private space, each augmented by more permeable outer perimeters of contorted paths, lights, motion detectors and video cameras projecting into the public realm of the sidewalk and the street. The public streets will become little more than interstitial space to these fortified private cores. They will themselves be fragmented by erecting barricading and monitoring by cameras overlapping each private space's permeable outer perimeter. Finally, overseeing it all will be helicopter patrols" (page 37).

In such a context, the worry is that future surveillant-simulation techniques will embed subjective normative assumptions about disciplining within cybernetic com­puterised systems of inclusion and exclusion, where even opportunities for human discretion are removed. Lyon (1994, page 211) asks the important question: "is what faces us a world of electronic technologies that classify us clinically, include and exclude by consumerist criteria, and are backed up by police and welfare depart­ments? ... will the new 'non-persons', segregated by surveillance systems, be failed consumers?" Surveillant-simulation techniques, bound together with new, fortified-cellular practices of producing built form, might well lead many urban areas towards the types of commodified, segregated enclosures already common in Los Angeles and Sao Paulo. Anything considered detrimental to the economic potential of a particular time-space zone, might be 'managed out' by combining digital biometric scan systems, wider geodemographic and social databases, and other planning, urban design, and policing practices. As Fyfe and Bannister (forthcoming) write, such a "fortress impulse", and the emergence of such "fortified cells", would increasingly link property values with "security", leading to a socially divisive process of "purification" which radically under­lines the ideal of urban public space. Norris et al (1996) warn that:

"those who cannot pay will be excluded from motorways; known trouble makers from football grounds; the unsightly casualties of 'care in the community' removed from the decorous order of city streets and shopping malls; known shoplifters and fare dodgers excluded from shops and transport systems.... If the growing divide between those who have and have not and those who are included and excluded is intensified through the use of new technology, there is a real danger that our cities will come to resemble the dystopian vision so beloved by futuristic film makers"

500 S Graham

Although I acknowledge the above trends, I also suggest that it is necessary to be extremely wary of simple black-white predictions of scenarios that assume totalised, dystopian, geographic 'impacts' of surveillant-simulation techniques. It is easy to assume the emergence of completely integrated, all-seeing surveillance webs when, in fact, the reality of technological innovation is a great deal less deterministic and a great deal more 'messy', difficult, contingent, and open to contested interpretations and applications (Bingham, 1996; Thrift, 1996). Surveillance, after all, is an essential ele­ment of all advanced societies, and surveillant-simulation techniques can bring bene­fits. Many individuals, for example, especially more affluent ones, might welcome the targeted, better quality services that such techniques might support (Lyon, 1994). Improvements for a wide range of socially useful services might be supported by surveillant-simulation techniques. BT, for example, are developing a 'smart home' application which 'learns' the activity patterns of housebound senior citizens and alerts care services if these patterns do not correspond to simulations of previous patterns. We must also remember that notions of privacy are culturally and socially contingent.

Above all, dispersed and networked systems of surveillant simulation do not amount to some omnipotent 'Big Brother'; rather, there exists a widening, deepening and broadening range of 'Little Brothers' (Lyon, 1994, page 53). In this context, Hannah argues that, although it is certain that "the modern citizen is objectified as a life-path comprised of information, as a 'spatialised dossier'" (1997, page 352), this dossier is far from becoming some omnipotent panopticon or all-seeing Big Brother. It is always incomplete, fragmented, and patchy, always partial, contingent, and unevenly developed across and between the 'life-paths' of citizens. Thus, "in 'real life' we face a variety of normalizing machines imperfectly coordinated, and each with imperfect powers" (Hannah, 1997, page 353). The lesson, if we are to understand the emerging practices of surveillant simulation, and the ways in which technologies become enrolled into the construction of space, is that we need to balance our macro political-economic treatments with much finer grained treatments of how technologies are socially con­structed to have effects in practice (Graham, 1998; Graham and Marvin, 1996).

Which brings us to our third question: what space exists for resistance strategies against the dominant spatial practices of surveillant simulation? Here, we need to explore the ambivalence of surveillance-simulation techniques, the degree to which their use is open to appropriation, and how community groups, activists, and those resisting dominant spatial practices, might utilise surveillant-simulation techniques to further their own practices. For space certainly remains in the interstices of technologies and material geographies to challenge these practices. As Kaliski (1994, page 42) argues:

"Curiously, many of the social transactions that are shaping the tenor of culture occur in the very places most subject to the scan of globalism. Shopping mall culture, gated enclaves (whether suburbs or rock houses), omnipresent surveillance and recording of everyday life do not seem to limit ever new and evolving cultural expressions and mutations born of unexpected gatherings. The easy reduction of these places to unitary theories or definitions of globalized spaces overlooks the physical workings of their quotidian elements."

Surveillant-simulation techniques, too, demonstrate such ambivalence. For example, GIS-based techniques can be as effective at revealing the emergence of fragmented geographies as they are at perpetuating them. Ramasubramian (1996) outlines how GIS techniques were used in Milwaukee to prove successfully that an insurance firm was starkly red-lining African-American census tracts in the city (ironically, by using its own GIS applications) (see also Harris et al, 1995). Robins (1996, page 139) has argued that, with the mass diffusion of consumer video, "the city now constitutes a mosaic of micro-visions and micro-visibilities. With the camcording of the city we

Spaces of survcillant simulation 501

have the fragmentation and devolution of vision-as-control to the individual level**. And although there appear to be few "revolts against the gaze" (Fyfe and Banister, 1997, page 13) of CCTV, at least in UK cities, Morris and Armstrong (1997) have shown how the young men under most scrutiny develop sophisticated strategies for exploiting its weak spots, so minimising its effects on them. Ambivalent relations emerge here. The Rodney King case demonstrated, for example, that cityvvide CCTV, if properly configured, regulated, and used, might substantially reduce the more obvious abuses of power by law enforcement agencies. Moreover, Shields (1994) has shown that even shopping malls may be appropriated as public spaces by marginalised groups, who struggle against intense scrutiny and surveillance to access and use it for their purposes.

Thus, attention needs to centre on the complex, ambivalent relations surrounding surveillant-simulation techniques whilst still being sensitive to overall macrolevel biases in their design, deployment, and operation, and to their linkages to the spatial prac­tices of dominant organisations. Boyer (1997), in fact, has suggested that those resisting dominant spatial practices need to construct their own simulations•- both virtual and material as a kind of 'radical artifice*, geared towards supporting a new associational urban politics which undermines the splintering of urban spaces into spaces of enclo­sure and exclusion.

Our final question centres on how surveillant-simulation techniques arc becoming implicated in the elaboration of new material geographies of employment, urbanisa­tion, and flow. For the proliferation of surveillance systems is about much more than Hows of representations; of the construction of virtualities and simulacra; of mecha­nisms for (attempted) fine-grained social control; and of cybernetic processes of automation. It also fuels some of the fastest growing economic sectors of the 'informa­tion economy' (sec Graham and Marvin, 1996, chapter 5; Hepworth, 1989). The eco­nomic (lows and labour processes surrounding the growth of surveillant-simulation systems are a key element within the complex articulation between urban and regional economies and networked disembodied flows and relations (Castells, 1997).

Thus, data warehousing and consumer marketing industries, for example, generate huge demands for sophisticated 'switched-hV office space located in places with good labour supplies, public subsidies, and adequate transport, telecommunications, and property infrastructures. Such back office and telesales zones tend to locate far from the main urban cores in lower cost urban, rural, or Third World' spaces (Graham and Marvin, 1996; Richardson, 1994). Lower level 'data crunching' functions are often outsourced to even more dispersed locations, employing female staff on pay-per-key-stroke wages (and, ironically, employing surveillant-simulation techniques to support worker discipline and performance). The customer-support infrastructures for utilities, telecommunications, and transport firms now routinely operate, on-line, from cheap, distant, automated call centres, far from the territorial 'patches' covered by their infra­structures. And the flows of images from CCTV systems can now be easily switched over broadband networks to cheap labour locations. The World Bank has even sug­gested that the CCTV systems covering US malls should be monitored in Africa, to take advantage of low wage costs and offer 'developmental' benefits to the continent.

Acknowledgements. Grateful thanks to the extremely useful comments of the referees who reviewed this paper and to Geraldine Pratt for her helpful suggestions All remaining errors, of course, remain the author's responsibility.

References Batey P, Brown P, 1995, "From human ecology to customer targeting: the evolution of

geodemographics", in GISfor Business and Service Planning Eds P Longley, G Clarke (Geoinformation International, Cambridge) pp 77-103

502 S Graham

Batty M, 1996, "Simulating reality" Environment and Planning B: Planning and Design 23 253-254 Batty M, 1997, "Urban systems as cellular automata" Environment and Planning B: Planning and

Design 24 159 -164 Baudrillard J, 1983 Simulations (Semiotext(e), New York) Baudrillard J, 1994 Simulacra and Simulation (Michigan State University Press, East Lansing, MI) Bingham N, 1996, "Object-ions: from technological determinism towards geographies of relations"

Environment and Planning D: Society and Space 14 635 - 657 Bleecker J, 1994, "Urban crisis: past, present, and virtual" Socialist Review 24 189 - 221 Bogard W, 1996 The Simulation of Surveillance: Hyper control in Telematic Societies (Cambridge

University Press, Cambridge) Boyer C, 1996 Cybercities: Visual Perception in the Age of Electronic Communication (Princeton

Architectural Press, New York) Boyer C, 1997, "Cybercities: virtual representations and global cities", paper presented to the

Global City-Regions conference, Madrid, July; copy available from the School of Architecture, Princeton University, NJ, USA

Bray P, 1995, "A map that fills in the gaps" The Guardian online supplement, 2 February, page 6 Burrows R, 1997a, "Virtual culture, urban social polarisation and social science fiction", in The

Governance of Cyberspace Ed. B Loader (Routledge, London) pp 38-45 Burrows R, 1997b, "Cyberpunk as social theory: William Gibson and the sociological imagination",

in Imaging Cities Eds J Williams, S Westwood (Routledge, London) pp 234-248 Burstein D, Kline D, 1995 Road Warriors: Dreams and Nightmares Along the Information Highway

(Dutton, New York) Castells M, 1997 The Rise of the Networked Society (Blackwell, Oxford) Crawford R, 1996, "Computer-assisted crises", in Invisible Crises: What Conglomerate Control of

Media Means for America and the World Eds G Gerbner, H Mowlana, H Schiller (Westview, Boulder, CO) pp 47-81

Davies S, 1996 Big Brother: Britain's Web of Surveillance and the New Technological Order (Pan Books, London)

Davis A, 1997, "The body as password" Wired July issue, pages 132- 140 Davis M, 1990 City of Quartz (Vintage, London) Davis M, 1992, "Beyond Blade Runner: urban control, the ecology of fear", Open Magazine

Pamphlet Series, Westfield, New Jersey Dedrick, 1995, "A consumption model for targeted electronic advertising" IEEE Multimedia Summer

issue, pages 41 - 49 De Landa M, 1991 War in the Age of Intelligent Machines (Zone Books, New York) Deleuze G, 1988, "Postscript on the societies of control" October part 59, pages 3 -7 Der Derian J, 1990, "The (s)pace of international relations: simulation, surveillance, and speed"

International Relations Quarterly 34 295-310 Driver F, 1984, "Power, space and the body: A critical reassessment of Foucault's Discipline and

Punish" Environment and Planning D: Society and Space 3 425 - 446 Droege P (Ed.), 1997 Intelligent Environments (North-Holland, Amsterdam) Druckrey T (Ed.), 1994a Electronic Culture: Technology and Visual Representation (Aperture,

New York) Druckrey T, 1994b, "Introduction", in Electronic Culture: Technology and Visual Representation

Ed. T Druckrey (Aperture, New York) pp 13-26 Dunn T, 1993, "The new enclosures", in Reading Rodney King, Reading Urban Uprising

Ed. R Gooding-Williams (Routledge, New York) pp 186-205 Flusty S, 1994 Building Paranoia: The Proliferation of Interdictory Space and the Erosion of Spatial

Justice (Los Angeles Forum for Architecture and Urban Design, West Hollywood, CA) Foucault M, 1977 Discipline and Punish (Pantheon, New York) Fyfe N, Bannister J, forthcoming, "The 'eyes on the street': closed circuit television surveillance

and the city", in Images of the Street: Representation, Experience and Control in Public Space Ed. N Fyfe (Routledge, London)

Garfinkel T, 1996, "Why driver privacy must be part of ITS", in Converging Infrastructures: Intelligent Transport and the National Information Infrastructure Eds L Branscomb, J Keller (MIT Press, Cambridge, MA) pp 324-340

Gelerntner D, 1991 Mirror Worlds: The Day Software Puts the Universe in a Shoebox: How It Will Happen and What It Will Mean (Oxford University Press, New York)

Gilbert N, Conte R, 1994 Artificial Societies: The Computer Simulation of Social Life (UCL Press, London)

Spaces of stirvcillunt simulation 503

Goss J, 1995,"Marketing the new marketing: the strategic discourse of geodemographic information ^y$im\\$VinGrotmd'Truth: The Social Implied 3 Pickles (Guilford, New York) pp 130 -170

Gottdciner M, 1997 The Thcming of America: Dreams, Visions and Commercial Spaces (Wcstvicw Press, Boulder, CO)

Gowdy V, 1994, "Alternatives to prison'* The Futurist issue 28, page 53 Graham S, 1998, "The end of geography or the explosion of place? conceptualising space, place and

information technology" Progress in Human Geography 22(2) Graham S, Marvin S, 1994, "Cherry picking and social dumping: British utilities in (he 1990s"

Utilities Policy 4(2) 113-119 Graham S, Marvin S, 1996 Telecommunications and the City; Electronic Spaces, Urban Places

(Routledge, London) Graham S, Brooks J, Hcery D, 1996, 'Towns on the television: closed circuit TV in British towns

and cities" local Government Studies 22(3) 3 - 27 Gregory D, 1994 Geographical'Imaginations(Blackwell, Oxford) Guy S, Marvin S, 1995, "Pathways to 'smarter' utility meters: the socio-tcchnical shaping of new

metering technologies", WP 56^ Department of Town and Country Planning, University of Newcastle upon Tync, Newcastle upon Tync

Guy S, Graham S, Marvin S, 1997, "Splintering networks: cities and technical networks in a global-local age" Urban Studies 34 191 -216

Hall S, 1993 Mapping the Next Millennium (Random House, New York) Hannah M, 1997, "Imperfect panopticism: Envisioning the construction of normal lives", in

Space and Social Theory Eds G Bcnko, U Stohmayer (Blackwell, Oxford) pp 344 - 359 Haraway D, 1991, "A manifesto for cyborgs: science, technology, and socialist-feminism in the late

twentieth century", in Simians, Cyborgs and Women: The Reinvention of Nature Ed, D Haraway (Routeldgc, New York) pp 149 -181

Harris T, Weinter D, Warner T, Levin R, 1995, "Pursuing social goals through participatory GIS: redressing South Africa's historical political ecology", in Ground Truth: The Social Implication of Geographic Information Systems Ed. J Pickles (Guilford, New York) pp 196-222

Hcpworth M, 1989 The Geography of the Information Economy (Belhavcn, London) Hcwson M, 1994, "Surveillance and the global political economy", in The Global Political Economy of

Communications Ed. E Comor (Macmillan, Basingstoke, Hants) pp 61-79 Hillis K, 1996, "A geography of the eye: the technologies of virtual reality", in The Cultures of

Internet (Sage, London) pp 70-98 Jacobscn R, 1994, "Virtual worlds capture spatial reality" GIS World December issue, Ed. R Shields,

pages 36-39 Kaliski, 1994, "Liberation and the naming of paranoid space", in Building Paranoia: The Proliferation

of Interdictory Space and the Erosion of SpatialJustice Ed. S Flusty (Los Angeles Forum for Architecture and Urban Design, West Hollywood, CA)

Kellner D (Ed.), 1994 Baudrillard: A Critical Reader (Blackwell, Oxford) Latour B, 1993 We Have Never Been Modern (Harvard University Press, Cambridge, MA) Laws G, 1994, "Oppression, knowledge and the built environment" Political Geography 13(1) 7-32 Leyshon A, Thrift N, 1995, "Geographies of financial exclusion: financial abandonment in Britain

and the United States" Transactions of the Institute of British Geographers, New Series 20 312 - 341 Lefebvre H, 1984 The Production of Space (Blackwell, Oxford) Lilly R, 1990, "Selling justice: electronic monitoring and the security industry" Justice Quarterly

9493-503 Lilley R, Knapper P, 1993, "The corrections-commercial complex" Crime and Delinquency 39(2)

pages 150 -166 Lyon D, 1994 The Electronic Eye: The Rise of Surveillance Society (University of Minnesota Press,

Minneapolis, MN) Mckie R, 1994, "Never mind the quality, just feel the collar" The Observer 13 November, page 1 Massey J, 1996, "Keeping the customer satisfied" Information Age July/August issue, pages 22-24 Mowshowvitz A, 1996, "Social control and the network marketplace", in Computers, Surveillance, and

Privacy Eds D Lyon, E Zuriek (University of Minnesota Press, Minneapolis, MN) pp 79 -103 Norris C, Amstrong G, 1997, "Categories of control: the social construction of suspicion and

intervention in CCTV systems", report to ESRC; copy available from the Department of Social Policy, University of Hull, Hull

Norris C, Moran J, Armstrong G, 1996, "Algorithmic surveillance—the future of automated visual surveillance", mimeo, Department of Social Policy, University of Hull, Hull

504 S Graham

Olalquiaga C, 1992 Megalopolis: Contemporary Cultural Sensibilities (University of Minnesota Press, Minneapolis, MN)

Philo C, 1992, "Foucault's geography" Environment and Planning D: Society and Space 10 137 -161 Pickles J (Ed.), 1995 Ground Truth: The Social Implication of Geographic Information Systems

(Guilford, New York) Pile S, Thrift N, 1995, "Mapping the subject", in Mapping the Subject: Geographies of Cultural

Transformation Eds S Pile, N Thrift (Routledge, London) pp 13-51 Ramasubramian L, 1996, "Building communities: GIS and participatory decision making"

Journal of Urban Technology 3(10) 67-79 Richardson R, 1994, "Back officing front office functions—organisational and locational

implications of new telemediated services", in Management of Information and Communication Technologies Eds W Melody, R Mansell (Economic and Social Research Council, London) pp 309-335

Robins K, 1996 Into the Image: Culture and Politics in the Field of Vision (Routledge, London) Robins K, Hepworth M, 1988, "Electronic spaces: new technologies and the future of cities"

Futures April issue, pages 155 -176 Sawchuck K, 1994, "Semiotics, cybernetics, and the ecstasy of marketing communication", in

Baudrillard: A Critical Reader Ed. D Kellner (Blackwell, Oxford) pp 89-118 Saeed S, 1993, "British Telecom uses GIS to gear investment to the marketplace" GIS Europe

February issue, pages 10-12 Schiller H, 1996 Information Inequality: The Deepening Social Crisis in America (Routledge,

New York) Sheth J, Sisodia R, 1993, "The information mall" Telecommunications Policy 17 376- 389 Shields R, 1994 Places on the Margin (Routledge, London) Smith A, 1990, "Survival techniques for the Nineties" Marketing 22 January, pages 20 - 22 Soja E, 1996 Thirdspace (Blackwell, Oxford) Sorkin M (Ed.), 1992 Variations on a Theme Park (Hill and Wang, New York) Squires J, 1994, "Private lines, secluded places: privacy as political possibility" Environment and

Planning D: Society and Space 12 387-401 Stone A R, 1994, "Preface", in Electronic Culture: Technology and Visual Representation

Ed. T Druckrey (Aperture, New York) pp 6-10 Sui D, 1994, "GIS and urban studies: positivism, post-positivism, and beyond" Urban Geography

15 258-278 Thrift N, 1996, "Inhuman geographies: landscapes of speed, light and power", in Spatial Formations

Ed. N Thrift (Sage, London) pp 256 - 310 Tuathail G, LukeT, 1994, "Present at the (disintegration: deterritorialization and reterritorialization

in the new wor(l)d order" Annals of the Association of American Geographers 84 381 -398 Virilio P, 1987, "The overexposed city" Zone 1(2) 14-31 Virilio P, 1994, "The third interval: a critical transition", in Rethinking Technologies

Ed. V Andermatt-Conley (University of Minnesota Press, Minneapolis, MN) pp 3 -10 Weiss R, 1996, "Commercial gene tests raise spectre of DNA discrimination" The Guardian 28 May,

page 10 WilkieT, 1996, "Genes 'R' Us", in Future Natural: Nature, Science, Culture Eds G Robertson,

M Mash, L Tickner, J Bird (Routledge, London) pp 33 -145 Winkler M, 1991, "Walking prisons: the developing technology of electronic controls" The Futurist

July -August, pages 34-36 Yurick S, 1985 Beyond Metatron: The Recording Angel (Semiotext(e), New York)

p © 1998 a Pion publication printed in Great Britain